Speed test matrix

Qwen3.5-9B-FP8-dynamic on RTX 3090 24GB

vllm · 0.32.0 · FP8 dynamic6 contexts (0–232K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 128K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 30, 2026, 04:26 UTC

Hardware

Type: Discrete GPU
GPU: RTX 3090
VRAM: 24 GB
CPU: Intel Core i7-9700K (8C)
RAM: 32 GB
OS: Ubuntu 24.04.5

Engine

Engine: vllm
Version: 0.32.0
Quantization: FP8 dynamic
Backend: cuda
Repository: github.com/vllm-project/vllm
Build: vllm/vllm-openai:v0.32.0-x86_64

Run

Served model: bench
Revision: main
Benchmark tool: llm-inference-bench v0.8.0
Decode test: Fixed duration
Prefill mode: Measured during decode warm-up
Time per measurement: 30 s
Max output: 8,192 tokens

Launch command

# Untuned out-of-the-box baseline. Spec decode: none
# 1x RTX 3090 24GB (Zagitta/gpu-unlock), 420 W limit, PCIe Gen3 x16 | Intel Core i7-9700K 8C | 32 GB DDR4 | Ubuntu 24.04.5, NVIDIA driver 580.178.04
vllm serve RedHatAI/Qwen3.5-9B-FP8-dynamic --max-model-len 245760 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code --language-model-only

Prefill

prompt tok/s by context length

8K context7,1051,153 ms TTFT

16K context6,7892,389 ms TTFT

32K context6,3425,092 ms TTFT

64K context5,64511,412 ms TTFT

128K context4,59528,005 ms TTFT

Sustained decode

total tok/s · rows are context length, columns are simultaneous requests

Context	1	2	4	8	16	32	64	128
0	73.437 ms TTFT	140.451 ms TTFT	274.184 ms TTFT	534.7135 ms TTFT	961.2248 ms TTFT	Over capacity	Over capacity	Over capacity
16K	70.1112 ms TTFT	130.3161 ms TTFT	250.9270 ms TTFT	464.4433 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
32K	67.3111 ms TTFT	121.6154 ms TTFT	230.7225 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
64K	62.4194 ms TTFT	107.3261 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
128K	54.3522 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
232K	44.9778 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity

Over capacity: not run because this many requests at this context would exceed the server's KV cache.

TTFT is measured after the benchmark warms the server's prompt cache, so it mostly excludes prefill.

All measurements

Latencies are medians. Total tok/s is output across all simultaneous requests.

Phase	Context	Requests at once	Status	Total tok/s	Per request tok/s	TTFT	Time between tokens	Average active requests	Completed requests	Output tokens	Measured for
Prefill	8K	1	Valid	7,105.0	—	1,153 ms	—	—	1	—	—
Prefill	16K	1	Valid	6,789.0	—	2,389 ms	—	—	1	—	—
Prefill	32K	1	Valid	6,342.0	—	5,092 ms	—	—	1	—	—
Prefill	64K	1	Valid	5,645.0	—	11,412 ms	—	—	1	—	—
Prefill	128K	1	Valid	4,595.0	—	28,005 ms	—	—	1	—	—
Sustained decode	0	1	Valid	73.4	73.5	37 ms	13.6 ms	1.0	0	2,202	30.0 s
Sustained decode	0	2	Valid	140.4	70.3	51 ms	14.2 ms	2.0	0	4,212	30.0 s
Sustained decode	0	4	Valid	274.1	68.7	84 ms	14.6 ms	4.0	0	8,224	30.0 s
Sustained decode	0	8	Valid	534.7	67.1	135 ms	14.9 ms	8.0	0	16,040	30.0 s
Sustained decode	0	16	Valid	961.2	60.4	248 ms	16.5 ms	16.0	0	28,832	30.0 s
Sustained decode	0	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	70.1	70.2	112 ms	14.3 ms	1.0	0	2,103	30.0 s
Sustained decode	16K	2	Valid	130.3	65.2	161 ms	15.3 ms	2.0	0	3,910	30.0 s
Sustained decode	16K	4	Valid	250.9	62.6	270 ms	16.0 ms	4.0	0	7,524	30.0 s
Sustained decode	16K	8	Valid	464.4	57.7	433 ms	17.3 ms	8.0	0	13,928	30.0 s
Sustained decode	16K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	1	Valid	67.3	67.4	111 ms	14.8 ms	1.0	0	2,020	30.0 s
Sustained decode	32K	2	Valid	121.6	60.8	154 ms	16.4 ms	2.0	0	3,648	30.0 s
Sustained decode	32K	4	Valid	230.7	57.7	225 ms	17.3 ms	4.0	0	6,920	30.0 s
Sustained decode	32K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	1	Valid	62.4	62.4	194 ms	16.0 ms	1.0	0	1,870	30.0 s
Sustained decode	64K	2	Valid	107.3	53.7	261 ms	18.6 ms	2.0	0	3,218	30.0 s
Sustained decode	64K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	1	Valid	54.3	54.5	522 ms	18.4 ms	1.0	0	1,629	30.0 s
Sustained decode	128K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	232K	1	Valid	44.9	45.1	778 ms	22.2 ms	1.0	0	1,348	30.0 s
Sustained decode	232K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	232K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	232K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	232K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	232K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	232K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	232K	128	Over capacity	—	—	—	—	—	—	—	—

Over capacity: not run because this many requests at this context would exceed the server's KV cache.